Target speeds
APPENDIX 3 COLD STARTING & IDLING AT -20°C
A basic driveability evaluation was carried out in terms of engine stability through driver assessment of engine speed variations or other events.
Figure A3-1 Flowchart for cold engine starting and idling performance at -20°C
A3.3 RESULTS
Results are presented and discussed in Section A3.3.1 to A3.3.4 comparing the Baseline E10-E (BLUE) fuel with the Step 2 E10-E (RED) fuel. These fuels have E70 values of 51.9% (BLUE) and 60.6% (RED), with other parameters held constant as much as possible.
In Section A3.3.5, the more limited tests on the E10-E Step 1 (PURPLE) fuel are presented for completeness.
A3.3.1 Engine speed and temperature
Figure A3-2 Engine speed at idle following start at -20°C
Results are shown for the first 400s of the test. All the vehicles started easily (<1.6s) on both fuels and followed the same profile of high initial idle speed which decreased steadily as the engine warmed up. Engine speed profiles were consistent between tests (the occasional spikes in the traces represent temporary noise in the measuring system). Where there was some evidence of test to test variation (Vehicles 3 and 6), it appeared randomly and there was no evidence of a fuel effect.
Vehicle 5
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 4
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 1
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 2
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 3
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 5
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 4
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 1
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 2
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 3
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Figure A3-3 Engine coolant temperature at idle following cold engine starting at -20°C
Vehicle 6
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 5
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 2
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 1
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 3
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 4
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 5
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 2
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 1
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 3
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 4
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Coolant thermocouples were located in the top hose of the test vehicles (i.e. the
‘cold’ side of the thermostat). In spite of this, recorded temperatures started to rise very shortly after engine start and had reached 25-50°C after 400s. Temperatures generally followed consistent profiles for each test. Vehicles 3 and 6 again showed a small degree of test-to-test variability but this was not related to the fuel type.
A3.3.2 Exhaust Emissions
Under these cold start and warm-up conditions, we may expect to see significant levels of unburned or partially burnt fuel in the exhaust gases, particularly in the period of open loop operation before the lambda sensor and catalyst have warmed sufficiently to bring them under control.
Figure A3-4 shows that HC emissions rose to an initial peak, followed by a slow decay as the engines warmed up. The peak of the profiles is truncated, because emissions were higher than the maximum limit (just below 6000ppm propane equivalent) that could be measured by the analyser. There was good consistency between the individual tests.
Differences between the fuels were variable and relatively small: the more volatile red fuel gave lower emissions on Vehicles 1 and 4, but higher emissions on Vehicles 2 and 3 while there was no clear difference in Vehicles 5 and 6.
Figure A3-4 Exhaust HC emissions (in g/sec) at idle following cold engine starting at -20°C
For the CO emissions, peak values ranged from 7% to 11% and were all within the range of the exhaust analyser. Results in g/second are shown in Figure A3-5.
There were clear differences between the fuels, with the Step 2 E10-E (RED) fuel giving higher emissions on all the test vehicles compared to the Baseline E10-E (BLUE) fuel. The difference was very small for Vehicle 5.
Figure A3-5 Exhaust CO emissions (in g/sec) at idle following cold engine starting at - 20°C
Vehicle 1
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 3
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 4
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 5
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 2
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 1
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 3
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 4
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 5
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 2
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 100 200 300 400
Elapsed Time [Sec]
A3.3.3 Air-Fuel Ratio (lambda)
Exhaust AFR was measured using a heated UEGO sensor specially fitted to the vehicle. Because it was heated, the sensor could monitor lambda from the beginning of the test, whereas the vehicle’s own sensor takes time to warm up and establish lambda control. The vehicle’s own oxygen sensor was monitored but it was not disturbed and controlled the vehicle in the normal way.
Figure A3-6 Exhaust lambda values at idle following cold engine starting at -20°C
A clear transition from open-loop operation to lambda control was seen for Vehicles 3, 4, and 5 while the other vehicles approached lambda=1 in a more ambiguous way. Vehicle 2 appeared to reach lambda control around 50s but remained slightly rich of lambda=1 throughout the 1200s test. Vehicle 6 appeared to trend lean after a period of lambda=1 operation. The fuel did not affect the time taken to reach lambda control.
During the open-loop period, which varied between vehicles from below 50 to 300s, measured exhaust lambda values were lower (richer) on the Step 2 E10-E (RED) fuel.
A3.3.4 Discussion of the lambda and emission measurements
These lambda values are measured in the exhaust, whereas what we really want to understand is what differences occur inside the combustion chamber when the fuel volatility changes. To better understand the UEGO results, the data from the first 120s of the test have been analysed in more detail.
The engine’s control system will inject a specified volume of fuel according to the engine’s instantaneous needs. Since the BLUE and RED test fuels differ in energy content by only 0.12% per litre and by 0.02% per mg, we would expect the amount of fuel passing through the engine to be independent of which of the two fuels was used. Figure A3-7 shows that this is indeed the case. The fuel flow is calculated from the exhaust emission measurements, taking full account of unburned hydrocarbons and CO emissions.
Vehicle 4
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 5
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 3
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 1
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 2
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 4
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 5
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 3
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 1
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 2
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Small amounts of unburned fuel can enter the engine crankcase and over time dilute the lubricating oil. Apart from this, because gasoline is a light and volatile fuel, we do not expect any unburned fuel to remain in the combustion chamber or exhaust system, so it should all reach the exhaust analysers. Figure A3-7 suggests that any fuel passing to the crankcase is not significant in the overall fuel balance, and no difference between the fuels can be detected.
Figure A3-7 Calculated fuel flow in the first 120s after starting
During open-loop operation, the ECU will calculate the fuel demand based on engine speed and either mass air flow or manifold pressure. Direct measurements of mass air flow from the CAN system on Vehicles 1 and 2 (the only two vehicles where it was possible) confirmed this to be the case. Throttle positions and spark advance were measured for all vehicles except Vehicle 5 and showed no measureable differences between tests. Under these conditions, the amount of air drawn through the engine should be identical for the two fuels.
How can we then understand the exhaust lambda measurements, which showed differences between the fuels?
If the amount of fuel injected is the same for the two fuels, and the air flow the same, then the overall air-fuel ratio or lambda in the combustion chamber should also be the same.
Figure A3-8 shows emissions of CO, HC and CO2 during the first 120s for Vehicles 1 and 4. The increased CO emissions for the Step 2 E10-E (RED) fuel can be clearly seen, especially during the period from 20 to 60s after starting. At the peak, the difference is about 0.2 g/sec for Vehicle 1 and 0.1 g/sec for Vehicle 4.
For these vehicles, the RED fuel produced slightly lower HC emissions but the difference is only about 0.01g/sec, enough to produce about 0.02g/sec of CO, so this alone cannot explain the higher CO emissions. The CO2 emission traces show lower emissions on the RED fuel of 0.1g/sec for Vehicle 1 and 0.2 g/sec for Vehicle 4. While the precision of the measurements does not support a detailed analysis, it appears that combustion of the RED fuel is less complete, with more fuel carbon being emitted as CO rather than being completely converted to CO2.
For Vehicle 1, which shows the greatest difference, the amount of fuel carbon converted to CO rather than to CO2 may be as much as 10%, which would imply a 3% difference in energy output between the fuels. This is to some extent balanced
Vehicle 4: Fuel Cons (g) 10-120 seconds
0 10 20 30 40 50
RED BLUE BLUE RED RED
Vehicle 4 - Calculated Fuel mg/sec
0 100 200 300 400 500 600 700
0 20 40 60 80 100 120
Vehicle 4: Fuel Cons (g) 10-120 seconds
0 10 20 30 40 50
RED BLUE BLUE RED RED
Vehicle 4 - Calculated Fuel mg/sec
0 100 200 300 400 500 600 700
0 20 40 60 80 100 120
by the lower unburned hydrocarbon emission for the RED fuel (about 1%), but it is nevertheless surprising that the difference in energy output does not seem to affect the throttle position or manifold air flow needed to maintain the idle speed required by the closed-loop control. One possible explanation could be that less fuel is retained in the combustion chamber and lube oil at higher volatility but a more complete study would be needed to investigate this.
Figure A3-8 Exhaust CO, HC, and CO2 measurement in the first 120s after starting
This leaves the question of how we should interpret the exhaust lambda differences between the fuels. How the UEGO works is explained in [23]. The UEGO contains a zirconia measuring cell, and the air-fuel ratio is calculated by measuring the amount of oxygen that has to be electrochemically pumped in or out to achieve stoichiometry.
Under rich conditions, the air-fuel ratio is proportional to the amount of oxygen required to complete the oxidation of the unburned species (HC, CO and H2). In our tests, since the RED fuel is less completely burned than the BLUE fuel, the RED fuel should require more oxygen input to completely combust and hence be recorded by the UEGO as richer than the BLUE fuel. However, the diffusion rates of HC, CO and H2 are different and can affect the results. CO and HC tend to bias the sensor lean, while H2 (which we did not measure) biases it rich. The sensor
calibration is typically based on an artificial gas mixture with fixed proportions of CO, CO2, H2 and water so cannot adjust for compositional variations at the same AFR.
If we consider the conditions in the combustion chamber, we can say that there is no clear evidence that the more volatile fuel does have an overall richer mixture.
However, the distribution of fuel in the chamber will be different as evidenced by the compositional differences in the exhaust gases. The Step 2 E10-E (RED) fuel will evaporate more quickly and hence mix more thoroughly than the Baseline E10-E (BLUE) fuel. We cannot clearly deduce the conditions around the spark plug or at the cylinder wall, although better evaporation and mixing might be expected to reduce the degree of coking and deposit formation. The overall energy balance might suggest a higher lube oil dilution for the RED fuel, but a more extensive test would be needed to evaluate this.
In summary:
The fuel flow and air flow through the engines are the same for both fuels, which implies that the air-fuel ratio and lambda should be the same. However the exhaust UEGO measurement calculates a different lambda for the two fuels.
The RED fuel produces more CO and less CO2 in the exhaust than the BLUE fuel, while unburned hydrocarbon levels are slightly lower on the RED fuel. This should require a higher fuel flow for the RED fuel to sustain idle speed but this does not seem to be the case in reality.
The exhaust UEGO sensor output should be interpreted with caution. It is not surprising that the UEGO records a richer mixture for the RED fuel, because it is less completely burned than the BLUE fuel. However, the UEGO calibration is unlikely to compensate fully for compositional differences in the exhaust due to fuel changes and these may account for some of the differences seen.
Although we cannot directly measure conditions in the combustion chamber, we can deduce that the more volatile fuel gives better evaporation and mixing. It is not clear whether the effects of this are beneficial or detrimental.
A3.3.5 Results on Step 1 E10-E (PURPLE) fuel
The Step 1 E10-E (PURPLE) fuel was tested on Vehicles 5 and 6. Since these tests were performed after the main test series, repeat tests on the Baseline E10-E (BLUE) fuel were also performed. Key results are presented here for completeness.
These fuels have E70 values of 51.9% (BLUE) and 54.9% (PURPLE), with other parameters held constant as much as possible. Differences between the fuels are smaller than those for the Step 2 E10-E (RED) fuel, in line with the smaller change in E70.
Figure A3-9 Results for Vehicles 5 and 6 on the Step 1 E10-E (PURPLE) fuel
Vehicle 5
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Coolant Temperature °C
Vehicle 6
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 5
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Engine rpm
Vehicle 5
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
‐40
‐20 0 20 40 60 80
0 100 200 300 400
Elapsed Time [Sec]
Coolant Temperature °C
Vehicle 6
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 5
0 500 1000 1500 2000
0 100 200 300 400
Elapsed Time [Sec]
Engine rpm
Vehicle 5
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
0 0.2 0.4 0.6 0.8
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 5
0 0.1 0.2 0.3 0.4 0.5
0 100 200 300 400
Elapsed Time [Sec]
Lambda CO g/sec
Vehicle 5
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 6
0 0.2 0.4 0.6 0.8
0 100 200 300 400
Elapsed Time [Sec]
Vehicle 5
0 0.1 0.2 0.3 0.4 0.5
0 100 200 300 400
Elapsed Time [Sec]
Lambda CO g/sec